CN101206774B - Collation method and apparatus, registration apparatus and method - Google Patents
Collation method and apparatus, registration apparatus and method Download PDFInfo
- Publication number
- CN101206774B CN101206774B CN2007103016092A CN200710301609A CN101206774B CN 101206774 B CN101206774 B CN 101206774B CN 2007103016092 A CN2007103016092 A CN 2007103016092A CN 200710301609 A CN200710301609 A CN 200710301609A CN 101206774 B CN101206774 B CN 101206774B
- Authority
- CN
- China
- Prior art keywords
- registration
- similarity
- registration data
- checking
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
Abstract
The present invention provides a collation method which, with respect to plural registered data stored in a storage medium, calculates first degrees of similarity among the data before starting the collation, and, of the plural registered data, calculates a second degree of similarity between one registered data selected as the collation subject and reference data to be collated with the one registered data, and, in case the second degree of similarity is not a degree of similarity that can authorize that a person is a registrant, of other registered data other than the one registered data, excluding registered data which is collated as the collation subject in the past, of the first degrees of similarity among the one registered data and other registered data, selects one registered data having a degree of similarity closest to the second degree of similarity as the next collation subject.
Description
Technical field
The present invention relates to the expectation be applied to biometric authentication (biometrics authentication) check (collation) method, device and program are installed, are checked in registration (registration).
Background technology
Adopted such device, this device has been stored a plurality of people's that satisfy predetermined condition biology pattern therein as registration data (registered data), and registration data from a plurality of registration datas of such storage, selecting and and the reference data checked of this registration data between checked result (cross correlation value) be equal to or greater than under the situation of a threshold value personnel's approval be registration of personnel (registrant), the people who wherein satisfies predetermined condition for example is the people who is allowed to enter the people of the predetermined area or is allowed to use predetermined personal computer.
Generally speaking, select a such registration data in mode randomly continuously, till checked result (cross correlation value) is equal to or greater than predetermined value.
On the other hand, selecting in the situation of the collating sequence in a plurality of registration datas with random fashion, even personnel are registration of personnel, checking number of times also may be bigger.That is, be in the situation of registration of personnel personnel, treatment effeciency is unsettled.
In order to address this problem, proposed a kind of checking method, wherein, keep several reference pictures differ from one another and with first alphabetic data (its mode be by for every width of cloth registration image make similarity corresponding to registration image) of descending representative with the size of the similarity of registration image, and and check order related coefficient between second alphabetic data of similarity size of image based on first alphabetic data, a plurality of reference picture, with the descending representative, determine the collating sequence (for example, Japanese patent application discloses No.2006-139415 in early days) of registration image.
Summary of the invention
On the other hand, in this checking method, for check at every turn, must generate a plurality of reference pictures and check second alphabetic data of the similarity of image with descending representative, therefore, the number of registration image becomes big more, and the time period that is used to produce second alphabetic data just becomes long more.
Therefore, in this checking method, if surpass predetermined registration number, then with select a plurality of registration datas at random in collating sequence compare, may have such situation, what promptly treatment effeciency became in the situation that personnel are registration of personnel is even worse.
Consider above-mentioned situation, therefore be desirable to provide a kind of checking method, calling mechanism, check device and program, it is in the situation of registration of personnel personnel, select the situation of the collating sequence in a plurality of registration datas to compare at random with the number of not considering registration data, can improve treatment effeciency.
According to one embodiment of present invention, provide a kind of checking method, having comprised: first step, for a plurality of registration datas of storing in the storage medium, calculated first similarity between the described data before described the checking of beginning; Second step, in described a plurality of registration datas, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And third step, if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in described first similarity between a described registration data and other registration datas, a registration data selecting the most approaching described second similarity of similarity is as next checking object.
According to one embodiment of the invention, a kind of calling mechanism also is provided, comprise: computing unit, for a plurality of registration datas of storing in the storage medium, calculate when from described a plurality of registration datas, selecting a registration data as first similarity between the described data of choice criteria according to predetermined checking method; And storage control unit, storage medium is controlled, make that calculating described first similarity by described computing unit is stored in the described storage medium.
According to one embodiment of the invention, a kind of device of checking also is provided, comprise: computing unit, for a plurality of registration datas of storing in the storage medium, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And selected cell, if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in the described registration data and described first similarity between other registration datas in first similarity between a plurality of registration datas of storing in described storage medium, a registration data selecting the most approaching described second similarity of similarity is as next checking object.
According to one embodiment of the invention, also provide a kind of program that makes computing machine carry out following step: for a plurality of registration datas of storing in the storage medium, calculating when from described a plurality of registration datas, selecting a registration data according to predetermined checking method as first similarity between the described data of choice criteria, described storage medium is controlled, made described first similarity be stored in the described storage medium.
According to one aspect of the invention, a kind of program that makes computing machine carry out following step also is provided: for a plurality of registration datas of storing in the storage medium, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in the described registration data and described first similarity between other registration datas in first similarity between a plurality of registration datas of storing in described storage medium, a registration data selecting the most approaching described second similarity of similarity is as next checking object.
According to the present invention, if reference data is assumed that registration data, then because near with the selected next checking object of registration data of the similarity of reference data, so compare with the situation of collating sequence in selecting a plurality of registration datas at random, can improve and in the situation that personnel are registration of personnel, reduce the possibility of checking number of times.Even it is big that the number of registration data becomes, because no matter how the number of registration data does not calculate first similarity when checking, so can improve this possibility.
Therefore, such checking method, calling mechanism be can realize, device and program checked, compare with the situation of collating sequence in selecting a plurality of registration datas at random, how they can improve treatment effeciency regardless of the number of registration data in the situation that personnel are registration of personnel.
When read in conjunction with the accompanying drawings, can know essence of the present invention, principle and purposes, in the accompanying drawings, refer to similar part with similar label or character from following detailed.
Description of drawings
In the accompanying drawings:
Fig. 1 shows the synoptic diagram of the similarity degree between the registration data that uses when checking, be used to illustrate this checking method;
Fig. 2 shows the process flow diagram of this checking method of explanation;
Fig. 3 shows the synoptic diagram of illustrative experiment result (1);
Fig. 4 shows the synoptic diagram of illustrative experiment result (2);
Fig. 5 shows the synoptic diagram that the distribution of number of times is checked in explanation;
Fig. 6 shows the process flow diagram that method is determined in explanation;
Fig. 7 shows the synoptic diagram of Verification System according to an embodiment of the invention is described;
Fig. 8 shows the block diagram of the configuration of explanation calling mechanism;
Fig. 9 shows the process flow diagram of explanation registration process process;
Figure 10 shows the block diagram that the configuration of device is checked in explanation; And
Figure 11 shows the process flow diagram of explanation authentication processing process.
Embodiment
(1) checking method
With reference to figure 1 and Fig. 2 the checking method that utilizes registration data is described, supposes that in the method a plurality of registration datas are stored in the storage medium.In this checking method, in step SP1, before beginning to check,, obtain the similarity (hereinafter being called first similarity) between these data for a plurality of registration datas.
In Fig. 1, " T " represents registration data, and on behalf of registration number, appended numeral " X1, X2 ", appended numeral " 1 to n " represent middle registration number.In addition, the respective value among Fig. 1 is represented first similarity, and wherein on all four situation is set to " 1 ".In Fig. 1, for convenience, first similarity is according to registering series arrangement on matrix directions.
In the state that has obtained first similarity, will then be begun to check by data (hereinafter the being called reference data) R that checks with registration data if obtained.
That is, after having obtained reference data R, in this checking method, in step SP2, a plurality of registration data T
1To T
nIn X
2The inferior registration data Tx that is registered
2Be selected as initial checking object, and in step SP3, should elective registration data Tx
2R checks mutually with reference data.
Then, in this checking method, in step SP4, determine whether the similarity (hereinafter being called second similarity, is " 0.51 ") that obtains as checked result is to approve that personnel are similarities of registration of personnel in Fig. 1.
If having obtained is not to approve that personnel are the checked result of the similarity of registration of personnel (Fig. 2, step SP4 "No"), then in this checking method, in step SP5, at the registration data Tx as this checking object
2With other registration datas T
1To Tx
2-1, Tx
2+1To T
nSimilarity (X among Fig. 2
2The similarity of row) in, select with near the corresponding registration data Tx of similarity " 0.5 " of second similarity " 0.51 "
1As next checking object, turn back to step SP3 then, with registration data Tx
1R checks mutually with reference data.
As checked result, in step SP4, if registration data T
X1With second similarity of reference data R be not to approve that personnel are the similarity of registration of personnel (Fig. 2, step SP4 "No"), then in this checking method, in step SP5, get rid of the registration data Tx that past attempts is checked as checking object once more
2, just as the registration data T of checking object
X1With other registration datas T
1To Tx
1-1, Tx
1+1To T
nSimilarity (X among Fig. 2
1Similarity) in select with near the corresponding T of similarity " 0.5 " of second similarity " 0.51 "
1As next checking object.In having a plurality of and first similarity in the situation near the corresponding registration data of similarity of second similarity, select in these data any one as next checking object.
Like this, in this checking method, up to the checked result (second similarity) of registration data that is set as checking object and reference data be can approve that personnel are the similarity of registration of personnel before, in a plurality of registration datas behind the registration data that the eliminating past attempts was checked as checking object, in first similarity between the registration data that is set as checking object and other registration datas except that this registration data, select progressively has registration data near the similarity of second similarity as next checking object (Fig. 2, the circulation of step SP4-step SP5-step SP3).
Like this, in this checking method, in the supposition reference data is any one situation in a plurality of registration datas, owing to be selected as next checking object near registration data with the similarity of reference data, so compare with in a plurality of registration datas, selecting the situation of collating sequence at random, can greatly improve and in the situation that personnel are registration of personnel, reduce the possibility of checking number of times.
In addition, in this checking method, before beginning to check, obtain first similarity between a plurality of registration datas.Therefore, big even the number of registration data becomes in this checking method, because no matter the registration data number how, does not calculate first similarity, so can improve this possibility when checking.
Fig. 3 shows experimental result relevant with checking number of times in this checking method with Fig. 4.In Fig. 3 and Fig. 4, the image of the blood vessel pattern in 230 personnel's biosome zone is used as a plurality of registration datas.
In Fig. 3, be made as in the situation of initial checking object at registration data 230 personnel, will be except that being made as under the situation of reference data (data that are used to search for) as in other registration datas the registration data of initial checking object any one, the mean value of checking number of times when other corresponding registration datas are selected as second checking object be illustrated.
That is, for example first registration data is made as in the situation of initial checking object taking, in remaining second to the 230th registration data, for example, optionally the 115th registration data is made as reference data.Then, selected second checking object of corresponding second to the 230th registration data, and up to checked result (second similarity) be can approve that personnel are the similarity of registration of personnel before, according to this checking method, select progressively the 3rd and follow-up checking object.
In this example, be set as in the situation of the 115th registration data as second checking object of reference data in selection, checking number of times is secondary, and at second to the 230th registration data of selecting to have got rid of the 115th registration data during as second checking object, checks number of times and is three times or more times.In this example, first registration data is being made as in the situation of initial checking object, the mean value of checking number of times is 11 times.
From Fig. 3 as seen, even any registration data is made as incipient nucleus to value, can find out that also checking number of times is approximately 9 to 12 times.
On the other hand, in Fig. 4, multiple 10 registration data is set as under the situation of sum, in corresponding registration data is set as the situation of initial checking object, obtain to check the mean value of number of times, show the maximum kernel that becomes in the minimum situation at the mean value of checking number of times number of times and minimum are checked number of times.In Fig. 4, use " X " to show the mean value that minimum is checked number of times respectively, show with " Δ " and taking minimum to check in the situation of mean value of number of times maximum kernel, and use number of times
Show minimum and check number of times, and be shown in broken lines the mean value of checking number of times of working as in the situation of in a multiple registration data of 10, selecting collating sequence at random.
As shown in Figure 4,, select the situation of collating sequence to compare at random with registering in the image, also greatly reduced the mean value of checking number of times at 10 multiple even the number of registration image has increased.
On the other hand, add up to 230 registrations during images when noticing, from Fig. 5 as seen, it is 2 or 3 that great majority are checked number of times, compares with the mean value of checking number of times (9 times), may have the situation (maximum 60 times) that need much check number of times.That is, in second similarity is in situation in the scope that the distribution of first similarity between the registration data concentrates, because there are many similarities, so the nigh registration data of result is elected to be next checking object in proper order near second similarity.
Therefore, in checking the mean value of number of times, under the situation of maximum times less than designation number, when the registration data that will have the minimum average B configuration value before beginning to check is defined as being selected as registration data of initial checking object, compare with the situation of collating sequence in selecting a plurality of registration datas at random, when personnel are registration of personnel, can always reduce and check number of times.
(2) definite method of initial checking object
Particularly, with reference to the definite method that should be selected as a registration data of initial checking object of figure 6 explanations.
Determine method according to this, in step SP11, at a plurality of registration data T
1To T
nIn, select i (i=1,2 ..., or n) registration data T
iAs the initial candidate of checking object, and in step SP12, i registration data T of the initial candidate that is set as checking object will have been got rid of
iAfter other registration datas in any one registration data Tx (x=1,2 ..., or n, wherein i ≠ x) is defined as reference data.
Then, determine method, in step SP13, to SP5, obtain when selected second checking object of other corresponding registration datas, to check the maximal value and the mean value of number of times according to the step SP3 in the above-mentioned checking method according to this.
That is, for example, selecting the first registration data T
1In the situation as initial checking object, getting rid of registration data T
1Other registration datas T
2To T
nIn, the second registration data T
2Be selected as checking object.Then, with registration data T
2With the registration data T that is confirmed as reference data
xCheck mutually, and in checked result (second similarity) is not in the situation of the similarity that can approve that personnel are registration of personnel, before acquisition and the corresponding checked result of this similarity, step SP3 according to above-mentioned checking method arrives step SP5, by the checking object of select progressively the 3rd or back, obtain the number of times of checking when registration data T2 is elected to be second checking object.
In addition, at other registration datas T
2To T
nIn, for remaining registration data T
3To T
n, with registration data T
2Situation similar, obtain when with registration data T
3To T
nCheck number of times when being chosen as second checking object.Then, based on working as corresponding registration data T
2To T
nCheck number of times when being selected as second checking object, obtain as the first registration data T
1The maximum times of checking number of times and mean value when being selected as the candidate of initial checking object.
Like this, determine method, before in step SP14, obtaining positive result, be chosen in all registration data T according to this
1To T
nCheck the maximum times and the mean value of number of times when being selected as the candidate of initial checking object, then, in step SP15, in checking the mean value of number of times, under the condition of maximum times less than defined amount, the registration data that will be set as the candidate in the situation of getting the minimum average B configuration value is defined as initial checking object.
Like this, when checking, suppressed maximum number, compared, can always reduce personnel and check number of times during for registration of personnel with the situation of collating sequence in selecting a plurality of registration datas at random.
On the other hand, adopt the method for to determine,, can suppress maximum times, also may exist and to check the situation that just finishes through maximum times although compare with not adopting this situation of determining method.This is because registration data is selecteed under the collating sequence according to pre-defined rule (the step SP3 in the above-mentioned checking method is to SP5).
(3) change of checking method
Therefore, check number of times if on certain program, passed through certain, in the situation of this pre-defined rule of cancellation, promptly, pre-determined number selected in the situation of next checking object, when choosing at random the checking object of having got rid of the registration data that past attempts checked as checking object, compare with the situation of not cancelling this pre-defined rule, can improve and reduce the possibility of checking number of times that reaches before the target registration data.
Particularly, after the step SP4 in above-mentioned checking method, following step is set: determine to select the number of times of next checking object whether to surpass stipulated number, if and surpassed stipulated number, it then not the registration data that is selected as next checking object, but get rid of the registration data that past attempts was checked as checking object, choose at random a registration data.
The stipulated number of hope in this determining step is average time or less than checking number of times, and because the modal number of times of checking is 2 to 3 times (Fig. 5), so wish it is " 3 ".
Like this, if pre-determined number selected next checking object, when choosing at random the checking object of having got rid of the registration data that past attempts checked as checking object, check number of times in the time of can suppressing to check.
(4) embodiment
Now, the employing embodiments of the invention will be described with reference to the accompanying drawings.
(4-1) overall arrangement of Verification System
Fig. 7 shows Verification System 1 according to an embodiment of the invention.This Verification System 1 comprises management server 2 and at a plurality of the door openers of locating to arrange 3
1To 3
N, they are connected to management server 2 by dedicated line.
The calling mechanism 10 of management server 2 is gathered the image of user's's (hereinafter being called registration of personnel) that will be registered blood vessel, and extracts the blood vessel pattern in the image that obtains as the image acquisition result.Then, the image that calling mechanism 10 will wherein extract the blood vessel pattern registers among the storage medium HD, and as registration data, described registration is by making registration data and the corresponding realization of data of representing registration number.
In addition, when registering the registration data of second or back, calling mechanism 10 obtains to be registered in first similarity between a plurality of registration datas among the storage medium HD at every turn, determines to be selected as a registration data of initial checking object then.Then, the data storage that calling mechanism 10 will be represented the data of first similarity and represent the registration number in the registration data that is defined as initial checking object is perhaps upgraded these data in storage medium HD.
The image of blood vessel pattern of blood vessel that has wherein extracted the user that will enter the room as the reference data from door opener 3
1, 3
2..., or 3
NWhat be imported into management server 2 checks device 20.If only registration data is stored among the storage medium HD, then check the cross correlation value that device 20 calculates between input reference data and the registration data.
On the other hand, if two or more a plurality of registration data are stored among the storage medium HD, then utilize the data of representing first similarity and the data of representing the registration number in the registration data that is confirmed as initial checking object, according to predetermined checking method, check device 20 and calculate cross correlation value between input reference datas and these two or the more a plurality of registration data.
In addition, if predesignate check during obtain the cross correlation value bigger than threshold value, then check the data that device 20 will approve that personnel are registration of personnel and send to the door opener 3 that has sent this reference data
1, 3
2..., or 3
NOn the other hand, if do not obtain cross correlation value, then check device 20 and will not approve that these personnel are that the data of registration of personnel send to the door opener 3 that has sent this reference data greater than predetermined threshold
1, 3
2..., or 3
N
Door opener 3
1, 3
2..., or 3
NCollection will enter the user's in room the image of blood vessel, and extracts the blood vessel pattern in the image that obtains as the image sampling result.Then, door opener 3
1, 3
2..., or 3
NThe image that wherein extracted the blood vessel pattern is checked device 20 as what the reference data sent to management server 2, and wait for that personnel of approval are data of registration of personnel or do not approve that personnel are data of registration of personnel.
If receive data that personnel are registration of personnel of approval from checking device 20, then door opener 3
1, 3
2..., or 3
NOpen door, and if receive and do not approve that personnel are data of registration of personnel from checking device 20, then door opener 3
1, 3
2..., or 3
NKeep the door locking.
Like this, Verification System 1 can only make registration of personnel open and close door.
(4-2) configuration of calling mechanism
Next, will the configuration of calling mechanism 10 be described.As shown in Figure 8, calling mechanism 10 comprises control module 30, is connected to the image acquisition units 31 and the graphics processing unit 32 of control module 30 by bus 33.
When the fill order COM1 of the pattern (hereinafter being called the blood vessel enrollment mode) of sending the blood vessel that is used to register registration of personnel from operating unit 34, control module 30 arbitrarily controls image acquisition units 31 according to the program among the ROM of being stored in and graphics processing unit 32 is carried out registration process.
In the present embodiment, the blood vessel that adopts finger is as the image acquisition object, and when finger is placed to the image capture position place, the near infrared light that shines finger near infrared ray light source 31b passes finger interior to be reflected therein and scattering, then by optical system 31e, blood vessel projection light as the projection blood vessel enters image acquisition element 31d.In the case, because the opto-electronic conversion of image acquisition element 31d, the picture signal S1 of generation blood vessel is to be provided to driver element 31a.
Analog/digital (A/D) conversion process of driver element 31a carries out image signal S1, and will send to graphics processing unit 32 as the view data D1 that result obtains.
Like this, image acquisition units 31 has been gathered the image of blood vessel under the control of control module 30.
Blood vessel pattern in the image of graphics processing unit 32 extraction view data D1, the view data D2 that will obtain then as the extraction result of blood vessel pattern
i(i=1,2 ..., or n) send to control module 30.
As the concrete extracting method that is used for the blood vessel pattern, graphics processing unit 32 cuts out a part that comprises the blood vessel pattern in the image, extracts the profile of the blood vessel pattern in the image that cuts out like this.Then, graphics processing unit 32 will wherein extract the image binaryzation of the profile of blood vessel pattern, and the blood vessel pattern in the image of such binaryzation is attenuated.
To utilize flowchart text shown in Figure 9 to be registered in the registration process of control module 30 in the situation among the storage medium HD below at the registration data of second registration data or back.
When operating unit 36 receives the fill order COM1 that is used for the blood vessel enrollment mode, in step SP21, control module 30 control image acquisition units 31 are gathered the image of blood vessels, and obtain the image (view data D1 (Fig. 8)) that generates as the result of the image acquisition of being undertaken by image acquisition units 31.
Next, in step SP22, control module 30 control graphics processing units 32, make the blood vessel pattern in the image be extracted, and in following step SP23, by image is stored among the storage medium HD as registration data, be registered in the graphics processing unit 32 image (the view data D2 that the extraction result as the blood vessel pattern obtains
i(Fig. 8)).
Next, in step SP24, for two or more registration images of storing among the storage medium HD, cross correlation value between control module 30 computational datas, as first similarity, and in step SP25, passed through with above-mentioned definite method (Fig. 6) in from step SP11 after the similar treatment step of the corresponding steps of step SP15, determine to be selected as the registration data of first checking object.
At last, in step SP26, control module 30 is first similarity that calculates among the step SP24 with the content update of the data of representative first similarity of storing among the storage medium HD, and the content update that representative is defined as the data of the registration number in the registration data of initial checking object is the registration number in the registration data of determining in step SP25, finishes registration is handled then.
Like this, when registering the registration data of second registration data or back, 30 pairs of control modules are represented the data of first similarity between a plurality of registration datas of registering among the storage medium HD and are represented the data of the registration number in the registration data that is confirmed as initial checking object to upgrade at every turn.
In the situation that first registration data is registered among the storage medium HD, control module 30 is only carried out registration process through the step SP21 that associates to the corresponding steps of step SP23.
(4-3) check the configuration of device
On the other hand, as shown in figure 10 (wherein identical label be attached to Fig. 8 in the corresponding part of part), check device 20 and comprise control module 40, and the graphics processing unit 32, graphics processing unit 41, display unit 42 and the audio output unit 43 that are connected to control module 40 by bus 44.
When as door opener 3
1, 3
2..., or 3
NIn image acquisition result's view data D10 when communications processor element 41 is presented, according to program stored among the ROM, control module 40 arbitrarily control graphics processing unit 32, communications processor element 41, display unit 42 and audio output unit 43 is carried out authentication processing.
Communications processor element 41 is carried out and the door opener 3 of having exported view data D10
1, 3
2..., or 3
NMutual authentication (Fig. 1), and at control module 40 and the door opener 3 of having carried out mutual authentication
1, 3
2..., or 3
NBetween send and receive various data.
The represented content of video data that provides from control module 40 is provided on display screen display unit 42, and audio output unit 43 is provided from loudspeaker by the represented content of voice data that provides from control module 40.
Next, the authentication processing of flowchart text control module 40 shown in Figure 11 will be utilized.
When the communication data D10 that receives from communications processor element 41, in step SP31,40 start times of control module measure, and then in step SP32, the blood vessel pattern that control graphics processing unit 32 extracts in the image, and the data (reference data) of the image of blood vessel pattern have wherein been extracted in acquisition.
Next, in step SP33, control module 40 selects to be confirmed as the registration data of initial checking object, and is in step SP34, similar with above-mentioned steps SP3 (Fig. 1) then, the registration data D2 that will select like this by simple crosscorrelation
iD10 checks with reference data, and is in step SP35, similar with above-mentioned steps SP4 (Fig. 1) then, judges whether the cross correlation value (second similarity) that obtains as checked result is equal to or greater than predetermined threshold.
If obtain cross correlation value (second similarity) less than predetermined threshold, then control module 40 does not approve that personnel are registration of personnel, and in following step SP36, similar with above-mentioned steps SP5 (Fig. 1), get rid of the registration data that past attempts was checked as checking object, be set as the registration data D2 of checking object
iWith with the cross correlation value of other registration datas in (that is, first similarity: for example, the analog value corresponding to the X1 among Fig. 2 in capable), select the most approaching registration data D2 that is set as checking object of cross correlation value
iAnd the registration data of the cross correlation value between the reference data D10 (that is second similarity :) corresponding to " 0.51 " among Fig. 2.
Then, in step SP37, control module 40 judges to select the number of times of next checking object whether to surpass stipulated number, if and surpassed stipulated number, then replace the registration data that is selected as next checking object, the any registration data of gravity treatment is checked data as the next one, and is selecting to advance to next step SP38 after time numerical value is set to initial value.
On the other hand, if the number of times of selecting next checking object surpasses stipulated number, any registration data of not gravity treatment then, after will selecting inferior numerical value to increase progressively " 1 ", control module 40 advances to next step SP39.In step SP38, control module turns back to step SP34 and repeats above-mentioned processing, during determining to have passed through predetermined checking till.
Like this, during checking, up to the cross correlation value (second similarity) between registration data that is set as checking object and the reference data reach can approve that personnel are the value of registration of personnel before, from a plurality of registration datas, get rid of the registration data that past attempts was checked as checking object, control module 40 is according to pre-defined rule (that is step SP36) or optionally selection should be as a registration data of next checking object.
On the other hand, if obtained to be equal to or greater than the cross correlation value (second similarity) of predetermined threshold before during process is checked, then personnel of control module 40 approvals are registration of personnel, and advance to step SP40, by the door opener 3 of output reference data D10
1, 3
2..., or 3
N(Fig. 1), open corresponding door, finish this authentication processing.
On the other hand, if through during checking but do not obtained to be equal to or greater than the cross correlation value (second similarity) of predetermined threshold, then control module 40 approval personnel are third parties, and advance to step SP41, by display unit 42 (Fig. 1) and audio output unit 43 (Fig. 1), visually or with these personnel of audible notification be not recognized as registration of personnel, finish this authentication processing then.
(5) other embodiment
In the above-described embodiments, the image that has adopted the blood vessel pattern is as the registration data that is stored in the storage medium, but the present invention is not limited to this, but can adopt the image of various biological patterns, and for example, fingerprint, cheilogramma and nerve, and also can adopt unique point, for example, the take-off point of biological pattern or edge point, the present invention also can be used to the secret number except that biological pattern etc.
In addition, in the above-described embodiments, cross correlation value is used as first similarity and second similarity, but the present invention is not limited to this, but can adopt other various standards of representing similarity.
In addition, in the above-described embodiments, if pre-determined number selected checking object, then get rid of the registration data that past attempts was checked as checking object, choose at random checking object.On the other hand, as substituting to this, in with the preset range of second similarity as standard, from the registration data and first similarity between other registration datas that are set as checking object at this constantly, be equal to or greater than a defined amount similarity near the registration data of second similarity if exist, then get rid of past attempts as checking the registration data that data were checked, can choose at random next checking object.
Particularly, in above-mentioned steps SP36 (Figure 11), from the registration data and first similarity between other registration datas that are set as checking object at this constantly, detect with second similarity as the registration data in the preset range of standard, and in step SP37, judge whether the number of detected like this registration data is equal to or greater than specified number or amount.
Then,, then in step SP38, get rid of the registration data that past attempts was checked as checking object, choose at random next checking object, turn back to step SP34 then, repeat above-mentioned processing if the number of registration data is equal to or greater than specified number or amount.On the other hand, if the number of registration data is less than specified number or amount, then in step SP36 in the detected registration data, select a registration data near second similarity as next checking object, turn back to step SP34 then and repeat above-mentioned processing.
Like this, if checking object is selected on pre-determined number ground, the situation that then chooses at random checking object with getting rid of the registration data that past attempts checked as checking object is similar, can improve and reduce the possibility of checking number of times that reaches before the target registration data.
In addition, in the above-described embodiments, when each registration data is registered, determine to be selected as a registration data of initial checking object, but the invention is not restricted to this, but can determine registration data when per 100 registration datas are registered, and the registration data of each defined amount can be determined registration data when being registered in the storage medium.
In addition, in the above-described embodiments, carry out registration process (Figure 10) and authentication processing (Figure 11) according to the program that expands to from ROM on the RAM, but, the invention is not restricted to this, but can carry out registration process and authentication processing according to program of installing or the program of downloading from the Internet from storage medium, described storage medium for example is compact disk (CD), digital universal disc (DVD), semiconductor memory etc.
In addition, in the above-described embodiments, for checking function and registration function, used single assembly respectively, but the invention is not restricted to this, but single assembly can perhaps can come the single assembly branch with image collecting function by outfit to check function and registration function, that is, can adopt various configurations according to using.
In addition, in the above-described embodiments, as the effect when personnel are recognized as registration of personnel, door is opened, but the invention is not restricted to this, but can adopt the release effect of confined function in advance, perhaps state is changed into the effect of upstate, perhaps other effects from down state.The difference that depends on effect can arbitrarily change the configuration of Verification System.
The present invention can be applied to biometric authentication.
It will be appreciated by those skilled in the art that and depend on design requirement or other factors, can carry out various modifications, combination, sub-portfolio and replacement, as long as they are in the scope of appended claims or its equivalent.
The application comprises and the relevant theme of Japanese patent application JP 2006-343588 that was submitted to the Japan special permission Room on Dec 20th, 2006, and the full content of this application is incorporated herein by reference.
Claims (14)
1. checking method comprises:
First step for a plurality of registration datas of storing in the storage medium, calculated first similarity between the described data before described the checking of beginning;
Second step, in described a plurality of registration datas, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And
Third step, if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in described first similarity between a described registration data and other registration datas, a registration data selecting the most approaching described second similarity of similarity is as next checking object, and turns back to described second step.
2. checking method according to claim 1, wherein
In described third step, in with the preset range of described second similarity as standard, in described first similarity between a described registration data and other registration datas, if the number of the similarity of the most approaching described second similarity is equal to or greater than a defined amount, then get rid of the registration data that past attempts was checked as checking object, choose at random next checking object.
3. checking method according to claim 1, wherein
In described third step, if next checking object selected predetermined times, then get rid of the registration data that past attempts was checked as checking object, choose at random next checking object.
4. checking method according to claim 1 also comprises:
Determining step, definite registration data that should be selected as initial checking object before described the checking of beginning, wherein
In described determining step, if corresponding a plurality of registration data is set as the candidate of described initial checking object, then will be except being made as described reference data as any other registration data the candidate's of initial checking object the registration data, obtain when described other corresponding registration datas are selected as second checking object, to check the maximum times and the mean value of number of times, and
In the corresponding mean value of checking number of times that obtains in described corresponding a plurality of registration datas are set as candidate's the situation of described initial checking object, the registration data that is set as described candidate in the described maximum times of checking number of times is got the situation of minimum average B configuration value during less than defined amount is confirmed as being selected as a registration data of described initial checking object.
5. checking method according to claim 4, wherein
In described determining step,
When the registration data of each defined amount is registered in the described storage medium, determine to be selected as a registration data of initial checking object.
6. calling mechanism comprises:
Computing unit for a plurality of registration datas of storing in the storage medium, calculates first similarity between described a plurality of registration datas that will use when selecting a registration data according to predetermined checking method from described a plurality of registration datas,
Storage control unit is controlled described storage medium, makes described first similarity that is calculated by described computing unit be stored in the described storage medium.
7. calling mechanism according to claim 6, wherein
In described checking method, the registration data that is selected as checking object and will by and the reference data checked of a described registration data between second similarity be can approve that personnel are the similarity of registration of personnel before, in described a plurality of registration datas, get rid of the registration data that past attempts was checked as described checking object, in described first similarity between a described registration data and other registration datas except that a described registration data, a registration data of the most approaching described second similarity of similarity is selected as next checking object.
8. calling mechanism according to claim 6 also comprises:
Determining unit determines should be selected as a registration data of initial checking object, wherein in described checking method
Described storage control unit is controlled described storage medium, makes the identifier and described first similarity of a registration data being determined in described determining unit be stored in the described storage medium.
9. calling mechanism according to claim 8, wherein
If corresponding a plurality of registration data is set as the candidate of described initial checking object, then will be except being made as the data that will be retrieved as any other registration data the candidate's of initial checking object the registration data, described determining unit obtains the maximum times of checking number of times and the mean value that quilt is checked in the described checking method when described other corresponding registration datas are selected as second checking object, and
In the corresponding mean value of checking number of times that obtains in described corresponding a plurality of registration datas are set as candidate's the situation of described initial checking object, the registration data that will be set as described candidate in the described maximum times of checking number of times is got the situation of minimum average B configuration value during less than defined amount is defined as being selected as a registration data of described initial checking object.
10. check device for one kind, comprising:
Computing unit, for a plurality of registration datas of storing in the storage medium, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And
Selected cell, if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in the described registration data and described first similarity between other registration datas in first similarity between a plurality of registration datas of storing in described storage medium, a registration data selecting the most approaching described second similarity of similarity is as next checking object.
11. the device of checking according to claim 10, wherein
In with the preset range of described second similarity as standard, in described first similarity between a described registration data and other registration datas, if the number of the registration data of the most approaching described second similarity of similarity is equal to or greater than a defined amount, then get rid of the registration data that past attempts was checked as checking object, and described selected cell chooses at random next checking object.
12. the device of checking according to claim 10, wherein
If predetermined times that next checking object is selected then get rid of the registration data that past attempts was checked as checking object, and described selected cell chooses at random next checking object.
13. a register method comprises:
For a plurality of registration datas of storing in the storage medium, calculate first similarity between described a plurality of registration datas that when from described a plurality of registration datas, selecting a registration data, will use according to predetermined checking method,
Described storage medium is controlled, made described first similarity be stored in the described storage medium.
14. a checking method comprises:
First step, for a plurality of registration datas of storing in the storage medium, calculate registration data being selected as checking object and will by and the reference data checked of a described registration data between second similarity; And
Second step, if described second similarity is not to approve that personnel are similarities of registration of personnel, then in other registration datas except a described registration data, get rid of the registration data that past attempts was checked as checking object, in the described registration data and described first similarity between other registration datas in first similarity between a plurality of registration datas of in described storage medium, storing, a registration data selecting the most approaching described second similarity of similarity is as next checking object, and turns back to described first step.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2006-343588 | 2006-12-20 | ||
| JP2006343588A JP5018071B2 (en) | 2006-12-20 | 2006-12-20 | Verification method, registration device, verification device, and program |
| JP2006343588 | 2006-12-20 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101206774A CN101206774A (en) | 2008-06-25 |
| CN101206774B true CN101206774B (en) | 2010-10-13 |
Family
ID=39544401
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2007103016092A Expired - Fee Related CN101206774B (en) | 2006-12-20 | 2007-12-20 | Collation method and apparatus, registration apparatus and method |
Country Status (3)
| Country | Link |
|---|---|
| US (1) | US20080154929A1 (en) |
| JP (1) | JP5018071B2 (en) |
| CN (1) | CN101206774B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102087644B (en) * | 2009-12-07 | 2012-08-29 | 上海邮电设计咨询研究院有限公司 | Method for checking circuit data between BSS domain and OSS domain |
| JP5671607B2 (en) | 2011-03-29 | 2015-02-18 | 富士通フロンテック株式会社 | Biometric authentication device, biometric authentication system, and biometric authentication method |
| CN103544754B (en) * | 2013-10-25 | 2016-01-06 | 中安消技术有限公司 | A kind of method for unlocking of multi-lock and device |
| CN106055946A (en) * | 2016-05-18 | 2016-10-26 | 成都芯软科技发展有限公司 | System and method for identity recognition |
| CN109544343A (en) * | 2018-11-30 | 2019-03-29 | 阿里巴巴集团控股有限公司 | A kind of two-way trade liquidation method and system based on billing data |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1320887A (en) * | 2000-02-23 | 2001-11-07 | 日本电气株式会社 | Biological characteristic data acceptance method |
| JP2001338297A (en) * | 2000-05-29 | 2001-12-07 | Oki Electric Ind Co Ltd | Person identifying device |
| WO2005122468A1 (en) * | 2004-06-08 | 2005-12-22 | Nec Corporation | Data communication method and system |
| CN1716278A (en) * | 2004-06-28 | 2006-01-04 | 富士通株式会社 | The register method of biometric authentication system, biometric authentication system and program |
| CN1758264A (en) * | 2004-10-08 | 2006-04-12 | 富士通株式会社 | Biological authentification system register method, biological authentification system and program thereof |
| CN1773525A (en) * | 2004-11-11 | 2006-05-17 | 日本电气英富醍株式会社 | Collation device and collation method for authenticating users |
| JP2006139415A (en) * | 2004-11-10 | 2006-06-01 | Sony Corp | Authentication device, authentication method, and program |
| CN101017530A (en) * | 2006-02-10 | 2007-08-15 | 索尼株式会社 | Registration apparatus, collation apparatus and extraction method |
| CN101038629A (en) * | 2006-03-14 | 2007-09-19 | 富士通株式会社 | Biometric authentication method and biometric authentication apparatus |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3053607B2 (en) * | 1998-04-08 | 2000-06-19 | 三菱電機株式会社 | Data collation method and apparatus |
| JP2002163652A (en) * | 2000-11-27 | 2002-06-07 | Mitsubishi Electric Corp | Method and device for collating data |
| JP2003051012A (en) * | 2001-08-03 | 2003-02-21 | Nec Corp | Method and device for authenticating user |
| TWI299471B (en) * | 2001-08-24 | 2008-08-01 | Toshiba Kk | Person recognition apparatus |
-
2006
- 2006-12-20 JP JP2006343588A patent/JP5018071B2/en not_active Expired - Fee Related
-
2007
- 2007-12-03 US US11/949,510 patent/US20080154929A1/en not_active Abandoned
- 2007-12-20 CN CN2007103016092A patent/CN101206774B/en not_active Expired - Fee Related
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1320887A (en) * | 2000-02-23 | 2001-11-07 | 日本电气株式会社 | Biological characteristic data acceptance method |
| JP2001338297A (en) * | 2000-05-29 | 2001-12-07 | Oki Electric Ind Co Ltd | Person identifying device |
| WO2005122468A1 (en) * | 2004-06-08 | 2005-12-22 | Nec Corporation | Data communication method and system |
| CN1716278A (en) * | 2004-06-28 | 2006-01-04 | 富士通株式会社 | The register method of biometric authentication system, biometric authentication system and program |
| CN1758264A (en) * | 2004-10-08 | 2006-04-12 | 富士通株式会社 | Biological authentification system register method, biological authentification system and program thereof |
| JP2006139415A (en) * | 2004-11-10 | 2006-06-01 | Sony Corp | Authentication device, authentication method, and program |
| CN1773525A (en) * | 2004-11-11 | 2006-05-17 | 日本电气英富醍株式会社 | Collation device and collation method for authenticating users |
| CN101017530A (en) * | 2006-02-10 | 2007-08-15 | 索尼株式会社 | Registration apparatus, collation apparatus and extraction method |
| CN101038629A (en) * | 2006-03-14 | 2007-09-19 | 富士通株式会社 | Biometric authentication method and biometric authentication apparatus |
Non-Patent Citations (1)
| Title |
|---|
| CN 1716278 A,全文. |
Also Published As
| Publication number | Publication date |
|---|---|
| JP5018071B2 (en) | 2012-09-05 |
| JP2008158592A (en) | 2008-07-10 |
| CN101206774A (en) | 2008-06-25 |
| US20080154929A1 (en) | 2008-06-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110008903B (en) | Face recognition method, device, system, storage medium and face payment method | |
| CN106682068B (en) | Method and apparatus for adaptively updating a registration database for user authentication | |
| CN107038405B (en) | Method and apparatus for recognizing object and method and apparatus for training recognition model | |
| EP2784710B1 (en) | Method and system for validating personalized account identifiers using biometric authentication and self-learning algorithms | |
| US10990658B2 (en) | Method and apparatus for verifying user using multiple biometric verifiers | |
| CN108573137B (en) | Fingerprint verification method and device | |
| US20190130411A1 (en) | Method and system for data processing | |
| JP2007052720A (en) | Information access method by biometrics authentication and information processing system by biometrics authentication | |
| CN101206774B (en) | Collation method and apparatus, registration apparatus and method | |
| JP2017522634A5 (en) | ||
| US20170124428A1 (en) | Authentication method and apparatus, and method and apparatus for training a recognizer | |
| CN105659243B (en) | The implicit registration and mark being apprised of | |
| CN110162948A (en) | Method and apparatus for executing combination attestation | |
| WO2017173640A1 (en) | Method and apparatus for recognizing individuals based on multi-mode biological recognition information | |
| JP2009086951A (en) | Registration device, registration method, authentication device and authentication method | |
| CN111539740A (en) | Payment method, device and equipment | |
| EP3706021A1 (en) | Authentication system, authentication device, authentication method, and program | |
| CN111382666A (en) | Device and method with user authentication | |
| CN107368722A (en) | Verification method, computer-readable recording medium, the mobile terminal of biometric image | |
| KR20180068097A (en) | Method and device to recognize user | |
| CN113826135A (en) | System and method for contactless authentication using voice recognition | |
| CN112801067B (en) | Method for detecting iris light spot and computing equipment | |
| US20200380104A1 (en) | Radar-Based Behaviometric User Authentication | |
| JP2007334659A (en) | Individual authentication system | |
| US20220172537A1 (en) | Device and method for user authentication using security card |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20101013 Termination date: 20131220 |